System Maintenance Notice

Please note that maintenance work is scheduled between 4 AM and 5 AM - CET. During this time, you may experience reduced performance. Thanks for your patience.

Publications

Journal Article Ahead of print

PubMed: 42167356 DOI: 10.1002/resp.70266

Few-Shot Lung Cancer Classification via Electronic Nose Using Large Language Models: A Multicentre Prospective Study.

Respirology

Authors: Meng-Rui Lee, Chien-Chi Huang, Joyce Yue Sun, Chang-Ru Lin, Wen-Yuan Lin, Nai-Hui Chi, Kea-Tiong Tang, Jann-Yuan Wang, Chao-Chi Ho, Jin-Yuan Shih, Chong-Jen Yu

Affiliations

  • Department of Internal Medicine, National Taiwan University Hospital, Taipei, Taiwan.
  • Department of Internal Medicine, National Taiwan University Hospital Hsin-Chu Branch, Hsinchu, Taiwan.
  • Department of Electrical Engineering, National Tsing Hua University, Hsinchu, Taiwan.
  • International Bilingual School at Hsinchu-Science-Park, Hsinchu, Taiwan.
  • Department of Internal Medicine, National Taiwan University Hospital, Taipei, Taiwan. Department of Internal Medicine, Min-Sheng General Hospital, Taoyuan, Taiwan.
  • Department of Nursing, National Taiwan University Hospital, Taipei, Taiwan.

Abstract

BACKGROUND AND OBJECTIVE

Electronic Nose (eNose) breathprints are promising non-invasive lung cancer diagnostic tools, but cross-site validation and adaptation remain barriers to clinical applications. It remains unknown whether a natural language processing-pretrained large language model (LLM) can enable few-shot, site-specific classification of lung cancer using eNose breathprints.

METHODS

We collected eNose breathprints of lung cancer and non-lung cancer patients from two medical centres in Taiwan. A GPT-2-backbone LLM with parameter-efficient adaptation was compared with convolutional neural networks (CNN) trained from scratch or pretrained on CIFAR-100. Few-shot protocols (2-6 shots per class) and full-data training were evaluated.

RESULTS

We collected 432 eNose breathprints from two sites (S1 and S2). With 6 labelled samples per class (6 shots), LLM achieved an area under the curve (AUC) of 0.79 (95% CI: 0.71-0.87), sensitivity of 0.74 (0.63-0.83), and specificity of 0.77 (0.67-0.87) on S1. On S2, it achieved an AUC of 0.76 (0.69-0.82), sensitivity of 0.77 (0.69-0.84), and specificity of 0.61 (0.51-0.70). LLM outperforms scratch CNN models (S1; AUC: 0.44, p = 0.0002) (S2; AUC: 0.63, p = 0.0198) and CNN pretrained on CIFAR-100 images (S1; AUC: 0.57, p = 0.0100) and (S2; AUC: 0.61, p = 0.0248). LLM or a CNN model trained on the source site fails to improve performance after transferring to the target site for fine-tuning; for the LLM, performance even deteriorates.

CONCLUSION

Our study demonstrates the potential of pretrained LLMs for few-shot lung cancer classification in a real-world mixed clinical cohort, reducing dependence on large training datasets.

MeSH

Language, Lung, Respiratory Tract Diseases, Neoplasms, Classification, Electronic Nose, Lung Neoplasms, Diagnosis, Nose, Prospective Studies